.png){kind=icon}
.png){kind=icon}
Return
Nabla fractional distributed aggregative optimization algorithm
.png){kind=icon}
PRE
AI
DOI：10.1016/j.jfranklin.2026.108522.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
A multitude of challenges within the Industrial Internet of Things (IIoT), characterized by their parallelism and privacy requirements, are addressed by formulating them as distributed optimization problems in multi-agent networks. A nabla fractional distributed optimization problem with an aggregation item in the IIoT is investigated. To handle the dependence of local cost functions on both local and global information, two auxiliary variables are introduced to estimate the aggregative function and the global gradient. To address the challenges in convergence proof posed by the non-locality and memory property of fractional calculus, the optimality and convergence performance of the algorithm are analyzed from both system and frequency-domain perspectives. Its convergence with a Mittag-Leffler rate is proven by means of methods based on Lyapunov functions and time-frequency domain analysis. Motivated by the summarization and generalization of the commonalities of existing algorithms, a fractional distributed aggregative optimization algorithm framework is proposed. Numerical simulations on multi-UAV target protection and warehouse location validate the performance of the proposed algorithm.
Keywords:
fractional calculus
distributed optimization
multi-agent networks
aggregative function
convergence analysis
Journal
J
IF:
4.2
Papers:
822
Citations:
0